Previously, we just did three checks with some longer waits in between.
This was insufficient to avoid races, as we saw in practice. Now we
have slightly changed that approach and require 10 times in a row
an empty queue to deduce it atcually is empty.
closes https://github.com/rsyslog/rsyslog/issues/688
Remove the use sizeof(char) or sizeof(uchar) in calculations for memory
allocation or string length. There are no known platforms for which
sizeof(char) or sizeof(uchar) is not 1, and c99 has defined sizeof(char)
to be 1 (section 6.5.3.4 of c99).
- vg-test caught a minor issue in impstats (fixed)
- placeholder created for other dynstats tests
- a new assertion in diag.sh which allows extracting a particular column and expecting sum-up to match given value
- imdiag supports injecting litteral messages now (test can send a fully formed syslog message through imdiag, its a backward compatible change in protocol, allows for richer things to be build while being fully backward compatible)
so far, we checked only the main queue size to become zero, ignoring
the sizes of action queues. For some tests, this caused racieness and
unreliability. Now, we check all queues. This should make matters
much more stable.
This lead to a "module not found" error message instead of
the to-be-expected "module does not support new style" error message.
That invalid error message could be quite misleading and actually stop
people from addressing the real problem (aka "go nuts" ;))
the imdiag implementation of "queue empty" was a bit racy, which
sometimes lead to too-early termination of rsyslogd and thus some
small memory leaks (which in turn showed up as problems during
make check). The current architecture is not able to 100% guard against
this, but the probleme probability has been reduced (and it showed to
improve in practice).
The capability has been added for module to specify that they do not
like being unloaded.
related bug tracker: http://bugzilla.adiscon.com/show_bug.cgi?id=222
Signed-off-by: Rainer Gerhards <rgerhards@adiscon.com>
support for enhancing probability of memory addressing failure by
using non-NULL default value for malloced memory (optional, only if
requested by configure option). This helps to track down some
otherwise undetected issues within the testbench and is expected
to be very useful in the future.
simplified and thus speeded up the queue engine, also fixed some
potential race conditions (in very unusual shutdown conditions)
along the way. The threading model has seriously changes, so there may
be some regressions.
NOTE: the code passed basic tests, but there is still more work
and testing to be done. This commit should be treated with care.
